Study On Temperature Optimization Control Of Aluminum Electrolyzer Based On Status Classification

The aluminum electrolysis process is a nonlinear,multivariable and complex industrial process,and it is difficult to accurately describe the production status of the aluminum electrolysis process.Normally,we use the current efficiency of electrolyzer as an indicator to judge whether the electrolyzer is operating properly,and keeping the electrolysis temperature in a proper range can effectively improve the current efficiency.The paper proposes an electrolytic temperature optimization control strategy based on the status classification.It models and optimizes the key parameters of electrolytic temperature in the aluminum electrolysis process,so that the current efficiency is maintained at a high level to achieve high efficiency and low energy consumption of aluminum electrolysis.The first part is the soft-sensing model of the electrolytic temperature.Complicated physical and chemical reactions are accompanied with the process of aluminum electrolysis,there are many parameters which affected the electrolysis temperature and various parameters influence each other.It is difficult to model the electrolysis temperature in the aluminum electrolysis process by the method of mechanism modeling.This paper proposes an online nuclear extreme learning machine(OL-KELM)to establish a soft-sensing model for electrolyzing cell electrolysis temperature.Using the improved fuzzy curve method for the input parameters of the soft-sensing model to solve the problems of variable coupling and difficult identification in modeling,and then use OL-KELM method to conduct electrophoresis tank electrolysis temperature soft modeling.The second part is the slot classification of the electrolytic cell.A complete electrolysis cell parameter control system should have certain abnormal cell analysis capabilities and auxiliary decision-making capabilities.When the slot condition in the production process fluctuates,the current slot condition is discriminated by the slot condition evaluation model,and different operation measures are taken for different slot condition categories.By analyzing the characteristics of the channel condition,this paper uses fuzzy clustering(FCM)to establish the evaluation model of the condition of the channel,aiming at the problem that the number of clusters in fuzzy clustering is unknown and it is easy to lead into local optimum,Combining the intelligent optimization algorithm and the traditional clustering algorithms,an adaptive fuzzy clustering method based on the fruit fly algorithm(FOA)has been proposed to identify the bath conditions.The third part is the optimization control strategy of electrolytic temperature.In the current aluminum industry,when the situation of the tank is abnormal,most of the circumstances still rely on the operator's field experience to adjust.This method of monitoring based on qualitative experience knowledge is dependent on the technician and the accuracy is not high.In order to adjust the operating parameters that affect the electrolysis temperature within the normal range to keep the electrolysis temperature optimal,establish an electrolysis tank electrolysis temperature optimization control model based on intelligent optimization algorithm,then solved under the constraints of actual production process conditions in order to obtain a set of optimal process operating parameters.The effectiveness of the method was verified by simulation experiments based on the actual production process data of aluminum electrolysis.The fourth part is the implementation of electrolytic temperature optimization control system based on DCS platform.Taking into account the high temperature and highly corrosive environment in which the electrolytic cell is located,it is difficult to verify the research results obtained in the actual electrolytic cell.Based on the distributed control system(DCS)experimental platform,the actual production process of aluminum electrolysis can be simulated to verify the electrolytic temperature optimization control strategy studied in this paper.The simulation results show the tank electrolysis temperature optimization control scheme based on cell classification which proposed in the previous article has certain feasibility.